R/index.KL.r

In clusterSim: Searching for Optimal Clustering Procedure for a Data Set

index.KL<-function(x,clall,d=NULL,centrotypes="centroids")
{
wgss<-function(x,cl,d,centrotypes)
{
  n <- length(cl)
  k <- max(cl)
  centers<-matrix(nrow=k,ncol=ncol(x))
  for(i in 1:k){
    if(centrotypes=="centroids"){
      centers[i,]<-apply(x[cl==i,],2,mean)
    }
    else{
      centers[i,]<-.medoid(x[cl==i,],d[cl==i,cl==i])
   }
  }
  withins <- rep(0, k)
  x <- (x - centers[cl,])^2
  for(i in 1:k){
     withins[i] <- sum(x[cl==i,])
  }
  sum(withins)
}

   	 if(sum(c("centroids","medoids")==centrotypes)==0)
      stop("Wrong centrotypes argument")
   	 if("medoids"==centrotypes && is.null(d))
      stop("For argument centrotypes = 'medoids' d cannot be null")
     if(!is.null(d)){
      if(!is.matrix(d)){
        d<-as.matrix(d)
      }
     row.names(d)<-row.names(x)
     }
  if(is.null(dim(x))){
    dim(x)<-c(length(x),1)
  }
	m<-ncol(x)   	   	
	g<-k <- max(clall[,2])
	abs((g-1)^(2/m)*wgss(x,clall[,1],d,centrotypes)-g^(2/m)*wgss(x,clall[,2],d,centrotypes))/abs((g)^(2/m)*wgss(x,clall[,2],d,centrotypes)-(g+1)^(2/m)*wgss(x,clall[,3],d,centrotypes))
}